Sarasota audit to review computer code

[eomer]

http://www.heraldtribune.com/apps/pbcs.dll/frontpage

Audit to review computer code

By CAROL E. LEE

[e-mail]

SARASOTA COUNTY -- The computer source code that tells touch-screen voting machines how to run will be analyzed in the next phase of a state audit to determine what, if anything, went wrong in the Nov. 7 election.

The source code analysis has not started yet, but it is already generating controversy in the contested Congressional District 13 race, in which Republican Vern Buchanan was certified the winner by 369 votes.

The state Division of Elections' top choice for heading the review is Florida State University associate computer science professor Alec Yasinsac, an outspoken Republican who has advocated paperless voting machines in the past.

Democrats and voting rights activists charge that Yasinsac is too partisan to conduct an objective investigation.

-snip-

More...

[kster]

1. K&R.nt

[philb]

2. Not all machines had problematic ballot designs or problematic code, important to

look at all the ones that had problems


1. to see how many machines had problematic (biased) ballot designs

2. to see how many had problematic programming

Looking at a machine that didn't have the problems will tell nothing about the ones that did.

[suziedemocrat]

3. That will be difficult to examine the actual code on the bad machines.

It will probably be executable code and therefore unreadable. They'll probably get the official source code from the company that made the machines. How can we be sure the code they analyze is the actual code being executed on the "bad" machines?

[InsultComicDog]

6. I guess they could

compile the source code provided and see if the executables match?

[w4rma]

8. CRC check. (nt)

[dlaliberte]

11. Recompile to check binaries, if you have source code.

That is an excellent idea. If you trust the compiler, and if you have the source code, then you can compile to binary and compare whether it is identical with the binary code in the machine (if you trust the compare program).

However, please note that if no hacked code is found after the election that doesn't mean it wasn't there on election day. Hacks can remove themselves completely, so this auditing process could possibly prove that there was (sloppy) hacking, but it can not prove that there was no hacking.

You need the actual verifiably working machines on election day to prove there is no hacking at that time. But even then, you need to verify every single vote - i.e. you need a paper trail at the very least, no question about it. But even then, having the voter compare the paper against what they thought they voted is fraught with potential for "mistakes" or oversights. There is simply no point in trusting electronic voting machines in the first place, not until we have an entirely different way of verifying everything about the hardware and code (which must be open source, signed and sealed), and the network, and every step of the counting process. We are so far away from such a system, if people in the future ever get there, they will remember us as still being in the digital dark ages.

[philb]

12. I don't think the machines have been impounded since election day

People likely had access to them to do whatever they wanted, as far as I know.
But I still think a full audit should be carried out.
If a machine had a lot of undervotes there was a reason, and its likely that the reason could be determined
if the machine has not been tampered with since election day. There were a lot of machines that were problematic.

[dlaliberte]

14. Likely to determine hack only if hack was sloppy

You said "If a machine had a lot of undervotes there was a reason, and its likely that the reason could be determined
if the machine has not been tampered with since election day." But the software in the machine can tamper with itself after election day, and therefore it could remove all evidence of the tampering. In other words, it is possible that such tampering will not be detectable.

But you did say it is "likely" rather than always possible to determine if there was tampering, so maybe you are implying that it is likely that such tampering was done in a sloppy way, by sloppy thinking enemies of democracy. You have a point there.

[w4rma]

7. Compiled code *is* readable. It can be run through a decompiler. It's just a lot harder to read.

You need to be a top notch programmer to do this type of work. Not just any progammer can do it.

[tbyg52]

9. And we should not need top-notch programmers,

or any programmers at all, to let us vote. PAPER BALLOTS!

[w4rma]

10. Exactly. (nt)

[SarasotaDem]

4. Hoping

They have yet to test the ACTUAL machines that were used in the worst
precincts.

A friend who attended the rally said that the election worker she talked to
remarked that the machines tested that were used in the election were
randomly selected.

Too much FOX in the hen house for me ... :(

[groovedaddy]

5. What did they actually "recount?"

When they did the recount, what are they actually counting? Is it nothing more than a retabulation of what each precinct reported?
I have had this question for awhile now - with no paper trail, what is acutally counted?

[philb]

13. There were biased ballot designs like in Sarasota used in other close Cong. races- such as this one

Virginia Voter Machine Problems.

a. On many ballots in heavily Democratic neighborhoods, Jim Webb’s name was cut off on the screen. The ballots say: "James H. (Jim)" with no Webb.

b. New reports that ballots in Essex County have Jim's name split on 2 pages. The "James H (Jim)" on one page, "Webb" on the next.

c. Reports of voting machines in Isle of White that do not provide a clear image of the ballot, making voting a challenge.

http://www.dailykos.com/story/2006/11/6/1717/68014

There were also a lot of other dirty tricks used in the Allen/Webb race


This ballot design problem also happened in other states

[philb]

15. Florida, Robo-Calls paid for by RNC used in 2 close Congressional races, including D13

Florida, Robo-Calls paid for by RNC used in 2 close Congressional races,
in District 13 Christine Jennings race
and in Dist 16 in Florida,

http://www.bradenton.com/mld/bradenton/15897649.htm